Research On The Jump Beta Of Chinese Stock Market Based On Jump Regression Model

The return of financial assets is not always smooth even in the almost continuous time.It may have a large discrete volatility which is called as "jump".Because of the existence of jumps,it is difficult to make allocation of assets and risk management.In view of the large-scale stocks jumps at the same time,we think that there is a close relationship between the jumps occur in individual stocks and market.This paper will study the jump dependence between two processes,individual stock and market,based on the jump regression model for the first time,then apply jump betas estimating from models to build portfolios.We select the five-minute high-frequency data of CSI 300 Index and its constituent stocks for research.The empirical investigation results show that market jumps have significant impact on jumps of individual stock.It means that there exists aggregate jump risk in Chinese stock market.We also find that jump beta has the time-varying characteristic.When the market is stable,the systemic jump risk tends to be fluctuating in a small range,but when the market goes up or down,the mean of jump betas exhibits sudden and heavy changes in its dynamics,and especially in the situation of market crash,systemic jump risk will increase a lot.For further research,considering the asymmetric influence of changes in market returns,this paper decomposes the basic jump regression model into two parts based on the sign of market jumps.We can estimate positive jump beta and negative jump beta according to extended models.These two betas measure the sensitivity of stock jump return to the positive market jumps and negative market jumps.Empirical study suggests that investors are more sensitive to the negative market jumps,and negative jump beta changes more drastically in time.The application concerns jump betas in regression,and we find that jump beta factors are effective and profitable.Based on this,this paper will recommend several investment strategies.Through back testing,jump beta,positive jump beta and negative jump beta all have strong ability to distinguish stocks,and the value of factors has negative impact to the future return of stocks.The test results show that when holding period is 3 months,the factor return of jump beta and negative jump beta is significantly positive.Due to the features of these factors.We find that long-short portfolio based on jump beta can achieve the annualized return of 15.64%from 2011 to 2019,and 11.67%for long strategy.Comparing to market,they all perform much better.